>claude ai is better than chatgpt>runway gen-3 is sora quality and it's already out with people using itso openai has just been surpassed right and that's why they're constantly trying to imply that they have some insane secret tech to keep their investors from leaving?
>>101220305Why is it called open when it's not open at all?
I asked claude what does being catholically married to something means, and it asked me to consult a religious leader.Llama got it right.
Claude 3.5 Sonnet is good at coding but the context window is so small it is unusable for actual coding tasks.
>>101220543You're not supposed to throw your entire codebase there anon
>>101220612I already select documents that are relevant, but after a few messages back and forth, Claude has forgotten the beginning of the conversion. I can fill 200k tokens in no time.
I pasted Arabic text (historical quote) in the Claude prompt and forgot to prefix it with "translate", and it reprimanded me, in Arabic, for getting "extremist" and "political".So I did it again with the "translate" prefix, and Claude did the same in English, but replaced "extremist" and "political" with "sensitive" and "controversial".But to be fair, it was good at getting the literal lingual meaning of smaller bits of the quote, but still missed double/contextual meanings, and its guesses for meanings behind the quotes was very christian-like in describing a relationship between man and god in purely personal and spiritual terms, despite the quote being Islamic.
>>101220476always use jailbreak moron
>>101220712That was me testing the AI in its natural setting, genius.There was no need to try to outsmart its retardations then feel smart about it.I wasn't looking for any info, since what I asked about, I already knew.
I don’t get why they can’t add an Artifact feature. Is it that difficult to offer that? Also, what are some fun things that you’ve done with Artifacts? I’ve only had it make basic web apps. I’m curious if there are other fun use cases.
>>101220921These are all just gimmicks, they finally have to publish a 1 million context window model like Google.
They still haven't released their voice feature lmao. OpenAI is as bad as google at this point, running on empty promises and vaporware. Too bad that Anthropic puts out shit that's still so censored.